Bees, associative learning, and the lack of a dopaminergic system edit
I've been reading lots of stuff about the ability of bees to remember stuff, and make decisions based on reward. I've been thinking about how this compares with the mmalian work on the dopaminergic reward system, and how bees seem to be able to make the same reward decisions with simpler neuroanatomy than mammals. However, the circuitry that bees are using probably underpins a valuation system that would come AFTER the dopaminergic system. It's entirely likely that this is present in mammals too, and is similarly simple.
Having said that, bees can do associative learning. Associative learning (arguably) involves biasing the selection of rewarded actions. It seems at first a little strange to develop a system that biases the selection of actions that produce an unexpected event, when a system to bias only rewarded actions already exists. At the moment I can think of two potential explanations.
- 1. Biasing the selection of actions that produces and unexpected event is part of a mechanism that learns cause and effect, potentially for a selection mechanism to make 'choices' on. This is a little bit different from just biasing behaviour. Although behaviour learned through associative learning isn't mindlessly repeated in most animals (except possibly ICSS)
- 2. The dopaminergic system does replicate a lot of the functions of this hypothetical valuation system, but perhaps it was too complex or risky to modify the existing system, or the existing system needed to co-exist. Many cortical functions could arguably be developed from subcortical functions, but instead they add redundancy, or possibly even cooperative functions that is above the combined abilities of the two systems independantly.